%0 Journal Article %T 分级警报在不同情况下对自动驾驶接管的影响
The Impact of Graded Alerts on Self-Driving Takeovers in Different Scenarios %A 黄润峰 %J Operations Research and Fuzziology %P 534-551 %@ 2163-1530 %D 2025 %I Hans Publishing %R 10.12677/orf.2025.151049 %X 3级自动驾驶通过减少驾驶员参与驾驶活动，有效地降低了事故发生的数量。当自动驾驶系统超出其限定的范围时，系统会发出接管请求(TOR)来提醒驾驶员及时接管车辆。现有的关于接管请求的研究侧重于视觉、听觉、触觉或多模态接管请求方式等角度，对于具体的接管请求内容的研究较少。作为补充，本研究旨在研究基于声音的分级警报在不同的驾驶环境、TOR提前期的情况下对接管性能和驾驶员生理、心理负荷的影响。总计招募了40名受试者，他们每个人都在模拟驾驶器中经历了18次接管事件。结果表明，分级警报显著影响驾驶员的生理、心理负荷，警报级别越高，生理和心理负荷越大；TORlt显著影响接管时间，TORlt越长，接管时间越短，接管性能越好；驾驶环境显著影响接管时间，驾驶环境越复杂，接管时间越长，但驾驶员的驾驶表现更差。在5 s TORlt下，警报级别越高，驾驶员的接管时间越短；在7 s TORlt下，警报级别越低，驾驶员的接管时间越短。这表明，在一定范围内，不同的警报级别适用于不同的TORlt，即在特定的TORlt下，相应的警报有助于帮助驾驶员快速接管车辆。这些发现可能有助于自动驾驶安全设计和未来驾驶研究的参考。
Level 3 autonomous driving effectively reduces the number of accidents by reducing driver involvement in driving activities. When the automated driving system exceeds its limited range, the system sends a takeover request (TOR) to remind the driver to take over the vehicle in time. Existing research on takeover requests focuses on perspectives such as visual, auditory, haptic, or multimodal takeover request modalities, with less research on specific takeover request contents. As a complementary approach, this study aimed to investigate the effects of audible-based hierarchical alerts on takeover performance and drivers’ raw and psychological loads under different driving environments and TOR lead times. A total of 40 subjects were recruited, each of whom experienced 18 takeover events in a driving simulator. The results showed that graded alerts significantly affected drivers’ physiological and psychological loads, with higher alert levels resulting in higher physiological and psychological loads; TORlt significantly affected takeover time, with longer TORlt resulting in shorter takeover time and better takeover performance; and the driving environment significantly affected takeover time, with the more complex driving environment resulting in longer takeover time, but poorer drivers’ driving performance. Under 5 s TORlt, the higher the alert level, the shorter the driver’s takeover time; under 7 s TORlt, the lower the alert level, the shorter the driver’s takeover time. This suggests that, within a certain range, different alert levels are applicable to different TORlt, i.e., under a specific TORlt, the corresponding alert helps the driver to take over the vehicle quickly. These findings may help to inform autonomous driving safety design and future driving research. %K 分级警报， %K 接管请求前置时间， %K 接管绩效， %K 生理、心理负荷
Graded Alerts %K Takeover Request Lead Time %K Takeover Performance %K Physical and Mental Load %U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=107954